1,721,288 research outputs found
Simple Nonlinear Dual Window Operator for Edge Detection
No full textWe propose a nonlinear edge detection technique based on a two-concentric-circular-window operator. We perform a preliminary selection of edge candidates using a standard gradient and use the dual-window operator to reveal edges as zero-crossing points of a simple difference function depending only on the minimum and maximum values in the two windows. Comparisons with other well-established techniques are reported in terms of visual appearance and computational efficiency. They show that detected edges are surely comparable with Canny's and Laplacian of Gaussian algorithms, with a noteworthy reduction in terms of computational load
Variable frame rate for very low bit-rate video coding
No full textMotion description of objects time evolution often wastes bit-rate even if computed at a low fixed frame rate. Bit-rate reduction can be achieved taking advantage of the inactivity periods: in such time intervals there is no need for motion coding, as an accurate description of the scene can be obtained only considering the frame difference information. We propose to estimate the motion between couples of frames of the sequence only when the detected activity is high. The possibility of achieving very low bit rate values resides in the ability of dealing with precise motion estimators. To this end we propose here use of a modified block matching algorithm (BMA), the sliced-BMA, applied on separate sub-sets of each macroblock. Experimental results with test video-sequences show a noteworthy gain together with object segmentation and tracking capabilitie
Content-Based Image Retrieval over the Web using Query by Sketch and Relevance Feedback
No full textThis paper investigates the combined use of query by sketch and relevance feedback as techniques to ease user interaction and improve retrieval effectiveness in content-based image retrieval over the World Wide Web. To substantiate our ideas we implemented DrawSearch, a prototype image retrieval by content system that uses color, shape and texture to index and retrieve images. The system avails of Java applets for query by sketch and uses relevance feedback to allow users dynamically refine queries
ReIOS: Reflective architecting in the internet of objects
No full textSelf-adaptive systems are modern applications in which the running software system should be able to react on its own, by dynamically adapting its behavior, for sustaining a required set of qualities of service, and dynamic changes in the context or in user requirements. They are typically involved in Future Internet development such as the Internet of Things where interoperability, flexibility, and adaptability are key requirements. Convergence of contents, services, things and networks seems to be the cornerstone to fullfil these requirements. We propose a reflective approach to provide a common abstraction for automating the deployment of component based applications in the Internet of Things environment. The proposed framework allows the design of heterogeneous, distributed, and adaptive applications built on the component based software engineering paradigm. The framework considers a metamodel instantiated in a Rest middleware properly modified for allowing different implementations by using reflective design patterns. We are currently working to refine the framework metamodel and to validate it in several implementation domains
Non-standard inference services for mobile computing: concept abduction via m-OODBMS
No full textThough increased potentialities of handheld devices allow to apply discovery approaches designed for the Semantic Web, care has to be paid in re-designing original frameworks and algorithms. The paper presents a revision of basic inference services leveraging object-oriented Database Management System to perform semantic matchmaking and provide logical explanations. OWL-DL Knowledge Bases have been properly migrated towards an object oriented version to enable reasoning by addressing proper queries to a mobile DB. The proposed framework has been implemented and tested: preliminary results have been reporte
Concept Abduction and Contraction in Semantic-based P2P Environments
No full textReasoning engines are largely used in resource discovery and matchmaking scenarios where, given a request, they are able to provide a list of compatible items arranged in relevance order. A significant added value is the possibility to explain match outcomes in order to obtain information for modifying or refining early queries. Though the feasibility of running logic-based reasoning tasks over various knowledge bases has been widely proven on fixed servers, it is a challenging subject to execute inference processes on handheld devices. The paper presents a revised lightweight version of abduction and contraction algorithms (going back to previous works) for matchmaking in Description Logics in mobile ad-hoc contexts. Implementation and tests have been carried out in a mobile P2P case study based on a simplified Bluetooth interaction paradigm
>
No full textContent Based image retrieval has recently become one of the most active research areas, due to the massive increase in the amount and complexity of digitized data being stored, transmitted and accessed. We present here a prototype implementation of DRAWSEARCH, an image retrieval by content system that uses color and shape (and texture in the near future) features to index and retrieve images. The system, currently being tested and improved, is designed to increase interactivity with users posing queries over the Internet and avails of a Java client for query by sketch. It also implements relevance feedback to allow users dynamically refine queries. Experiments show that the proposed approach can greatly reduce the user's effort to compose a query while capturing his/her information need with greater precision
Semantic indexing for image retrieval using description logics
No full textWe propose all approach based on description logics for the sematic indexing and retrival of images containing complex objects. We aim at providing a conventional content-based image retrieval system, which adopts low-level features With image analysis, the ability to deal with the complex stucture of objects in real images. Starting from a region based segmentation of images we provide a syntax for the description of compex objects as composition of simpler "basic" shapes. An extensional semantics allows to define reasioning services, such as retrieval, classification, and subsumption. A prototype system has been implemented to substantiate our ideas. First, encouraging, results are presented and discussed here
- …
